Gertrude Mary Cox | Biography

Gertrude Cox organized and directed several agencies dedicated to research and teaching in statistics. "By her missionary zeal, her organizational ability and her appreciation of the need for a practical approach to the statistical needs of agricultural, biological and medical research workers she did much to counter the confused mass of theory emanating from mathematical statisticians, particularly in the United States, who had little contact with scientific research," eulogized Frank Yates in the Journal of the Royal Statistical Society.

Cox was born on a farm near Dayton, Iowa, on January 13, 1900, to John William Allen and Emmaline (Maddy) Cox. After graduating from Perry (Iowa) High School in 1918, she devoted several years to social service and training for the role of deaconess in the Methodist Episcopal Church. She spent part of that time caring for children in a Montana orphanage.

In 1925, in preparation for advancement to superintendent of the orphanage, Cox entered Iowa State College in Ames. Although she took courses in psychology, sociology, and other topics that would advance her intended career, she majored in mathematics because of her talent for it. She graduated in 1929 and registered for graduate work under the direction of Professor George Snedecor, a proponent of Ronald A. Fisher's statistical methods. Cox and Fisher became friends when he worked at Iowa State during the summers of 1931 and 1936. In 1931, she earned Iowa State's first M.S. degree in statistics, and for the next two years Cox worked as a graduate assistant at the University of California, Berkeley, studying psychological statistics.

In 1933, Snedecor asked Cox to return to work at Iowa State's new Statistical Laboratory, where she built a reputation of expertise in experimental design. By 1939, Cox had become an assistant professor at Iowa State, although her teaching and consulting activities never allowed her time to write a doctoral dissertation. Eventually, in 1958, Cox was awarded an honorary Doctor of Science degree by Iowa State.

When Snedecor was asked to recommend nominees to head the new Department of Experimental Statistics being formed at North Carolina State (NCS) College's School of Agriculture, he showed his list to Cox, who asked why her name was not included. He then added a footnote to his letter: "Of course if you would consider a woman for this position I would recommend Gertrude Cox of my staff." She was hired in 1940, becoming the first woman to head a department at North Carolina State.

In 1944, Cox assumed additional duties as director of the NCS Institute of Statistics, which she had organized. By 1946, the University of North Carolina (UNC) joined the Institute, taking responsibility for teaching statistical theory while NCS provided courses in methodology. Cox saw the Institute's mission as developing strong statistical programs throughout the South, a vision described by colleagues as "spreading the gospel according to St. Gertrude." Always insistent that good statistical analysis depended on adequate computations, Cox unhesitatingly embraced the computer era. NCS became one of the first colleges in the country to install an IBM 650 computer, and statisticians in Cox's organizations developed powerful statistical software programs. Cox helped create the Biometric Society in 1947 and edited its journal Biometrics from 1947 to 1955. In 1949, she became the first female member of the International Statistical Institute, and was elected president of the American Statistical Association seven years later.

In 1950, Cox and her colleague William Cochran published Experimental Designswhich was intended to be a reference book for research workers with little technical knowledge. In fact, it became a widely-used textbook that Frank Yates described nearly 30 years later as "still the best practical book on the design and analysis of replicated experiments." In her own experimental design classes, Cox taught by focusing on specific examples gleaned from her years of consulting experience.

Although Cox made substantial contributions to the theory of statistics and experimental design, Richard Anderson wrote in Biographical Memoirs that her most valuable contribution to science was organizing and administering programs. She was exceptionally successful in generating financial support for research. For example, one large grant Cox obtained from the General Education Board established a revolving fund that supported fundamental statistical research for many years. Cox also played an integral role in planning what would become the Research Triangle Institute (RTI)for consulting and research, uniting the resources of NCS, UNC, and Duke University. In 1960, she retired from NCS and became the first director of RTI's Statistics Section.

Cox loved world travel, and during her lifetime she made 23 trips to various international destinations. After retiring a second time in 1965, she spent a year in Egypt establishing the University of Cairo's Institute of Statistics. On five different occasions, Cox worked on statistical assistance programs in Thailand. At the age of 76, she toured Alaska and the Yukon Territory by bus, train, and boat.

Although she received numerous honors, including her 1975 election to the National Academy of Sciences, Cox was particularly pleased with the dedication of the statistics building at North Carolina State University as "Cox Hall" in 1970 and the establishment by her former students of the $200,000 Gertrude M. Cox Fellowship Fund for outstanding students in statistics at NCS in 1977.

Cox died of leukemia on October 17, 1978, at Duke University Medical Center in Durham. During the preceding year, she had kept meticulous records of her treatment and response, making herself the subject of her final experiment.